Opamp single supply line level audio buffer (voltage follower) design help

[cincin]

I have 2 Sony TC-105 reel to reel tape decks that I want to arrange into a Frippertronic style tape echo.
The only outputs are speaker level, so I'm planning on adding a line-level output jack from the pre-amp, just after the volume pot.
This line out will be connected to an audio mixer, which will then feed the first tape deck and the output, etc.
(The units have line-level inputs, so I'm good there.)

From my initial research, I've read that I really should be buffering the signal on that new line-out, to help the signal transfer over the wire (preventing capacitance problems), alleviate the load on the pre-amp, and provide a low impedance input for the mixer.

So I'm looking to build 2 small self-contained buffer boxes. One to go between each of the tape decks' line-outs and the mixer.

I have no idea what I'm doing but I love to learn.

- I'd like to keep it to a single supply opamp setup for simplicity, to use less parts, and keep the size of the device down. But I could easily be talked out of that.

- I'd like for it to work from an inexpensive switching power supply wall-wart.

- I'd like it to be of reasonably high quality, signal-wise (low noise), without spending too much on exotic op-amps. Those boxes will be used for future audio projects, not jus the TC-105s.

I'm looking for some circuit / design / opamp recommendations.

I already have a reasonable selection of op-amps, so if one those if well suited that would be ideal:

AD627
AD642
AD797
AD96173
BA10324
CA55BE
CLC430
KIA4558
LF351
LF353
LF444
LM311
LM324
LM358
LM386
LM741
LM13700
LT1028
MC1490
NE5532
OP37
OPA2134
TL062
TL072
TLC2272

Thanks gang!
Attached is the TC-105 schematics, in case it's useful.

[magic]

Is there no possibility of installing the buffers inside tape decks and using their own power supply?

Generally, it's fairly simple stuff. Wire the chip for unity gain (I suppose that's what you want), for single supply add AC coupling on the input and output. Add 100Ω series resistance on the output to prevent the opamp from oscillating with cable capacitance.

I would use some audio opamp with output short circuit protection like 5532 or 2134. In a pinch, 4558 or 072 could be good enough for many applicaitons, but they have "phase reversal" near the negative rail which can prouduce ugly startup glitches in AC-coupled, single-supply circuits.

[Phil1977]

I´d also try to tap the power supply of the cassette deck.

The easiest way is probably to find 9V or 12V somewhere. Many opamp circuits want bipolar power supply, but in your case it should be more than sufficient to use a virtual ground by a 1:1 voltage divider of the supply voltage. Of course you need to decouple the in- and output by capacitors.

[PGPG]

- I'd like for it to work from an inexpensive switching power supply wall-wart.
- I'd like it to be of reasonably high quality, signal-wise (low noise),

These two points are difficult to reconcile, I think. Inexpensive switching power supply wall-wart is probably ideal source of noise so 'forget quality'.

[wasedadoc]

I would not even bother with an opamp. Emitter follower on collector of X3. Emitter resistor to 0 volt rail. Output capacitor and resistor to 0 volt rail.

[cincin]

Well my first thought was to make them stand-alone boxes so they can be-reused in other projects. But by putting them inside the tape decks I don't have to worry about power supply, case, and I can drop a connector and a cable from the equation. So it's starting to sound pretty good.

I'll have to read on emitter follower buffer. Will consider.

I was doodling something from the Unity Gain layout on the TL072 datasheet, but I'll look for a 5532 based design to start from instead, if it's a better choice.

Cheers!

[PGPG]

I'll have to read on emitter follower buffer. Will consider.

Emitter follower has a gain of about 0.99 what practically is close enough to 1.

I was doodling something from the Unity Gain layout on the TL072 datasheet

Search information and read about TIM:
https://www.analog.com/en/resources/glossary/transient-intermodulation-distortion-tim.html

This are disturbances that are not detected by standard tests where you use sinusoidal input, but people hear it.
It happens when you use high gain amplifier with its gain reduced by external loop. For high switching signal amplifier at first reacts to them with its internal gain and after some (very short) time feedback tells it that it changed his output too much and should reduce it.
Operation amplifier has a DC gain in a range of 100 000 and in the Unity Gain layout it is reduced to 1 by feedback.

You have 100% that emitter follower will not have TIM disturbances.

[cincin]

Apparently, one disadvantage of an emitter follower setup is that it's more sensitive to temperature changes.
If I put the circuit inside the reel to reel decks, could that be a problem?
There's a decent size AC motor in there, not to mention the power amp section from the 60's... I haven't specifically done operating temperature testing, but I would guess there's a fair bit of variance.

[wasedadoc]

Apparently, one disadvantage of an emitter follower setup is that it's more sensitive to temperature changes.
If I put the circuit inside the reel to reel decks, could that be a problem?
There's a decent size AC motor in there, not to mention the power amp section from the 60's... I haven't specifically done operating temperature testing, but I would guess there's a fair bit of variance.

Yes, the dc shift of an emitter follower depends on the base-emitter voltage of the transistor and that does change slightly with temperature.  However that is irrelevant here as the dc bias point (of the transistor - nothing to do with AC bias to the head) is not critical and the output from the emitter will be ac coupled through a capacitor.

[cincin]

Alright,

I'm looking at this circuit, using a 2N3904.

Seems to work, but there's a slight negative DC bias on the output, and I'm not sure why?

[MarkT]

If you are taking the feed from the wiper of the volume control you should be using a JFET or JFET opamp - don't want any appreciable bias current from the wiper or it'll make horrendous sounds on the slightest scratchiness...

I'd go with opamp - an emitter follower can only drop the impedance by a factor equal to the gain, perhaps 100 or so, so if the volume pot is 10k that's a buffer with perhaps 50 ohms output impedance.  Emitter follower is also rather non-linear unless loaded with a constant current source.  Opamp is simpler, just remember to decouple it.  For instance OPA1652 is nearly rail-to-rail, JFET, low distortion, low noise.

[wasedadoc]

Alright,

I'm looking at this circuit, using a 2N3904.

Seems to work, but there's a slight negative DC bias on the output, and I'm not sure why?

In actual circuit you do not need the base resistors. Base of NPN directly to collector of X3. Use 2k2 emitter resistor. 22uF capacitor, positive end to emitter. 10k resistor from negative end to 0 Volt.

[wasedadoc]

If you are taking the feed from the wiper of the volume control you should be using a JFET or JFET opamp - don't want any appreciable bias current from the wiper or it'll make horrendous sounds on the slightest scratchiness...

I'd go with opamp - an emitter follower can only drop the impedance by a factor equal to the gain, perhaps 100 or so, so if the volume pot is 10k that's a buffer with perhaps 50 ohms output impedance.  Emitter follower is also rather non-linear unless loaded with a constant current source.  Opamp is simpler, just remember to decouple it.  For instance OPA1652 is nearly rail-to-rail, JFET, low distortion, low noise.

This is not and never will be a low distortion system.  An analogue tape machine of this calibre is not going to get better than about 2%. And it has an audio transformer in the recording path. Rail to rail opamp is no benefit as the signal is not rail to rail. None of the preceding transistor stages are designed for rail to rail.

It is simple to plonk in the 4 components I suggest and try it by listening

[cincin]

In actual circuit you do not need the base resistors. Base of NPN directly to collector of X3. Use 2k2 emitter resistor. 22uF capacitor, positive end to emitter. 10k resistor from negative end to 0 Volt.

I'm afraid I don't really understand what you mean.
Do you mean I don't need to add DC bias to the signal via the voltage divider of the the two 100K resistors connected to the base of the transistor?
What is X3?

Sorry, I'm still quite new to all of this.

[cincin]

I'll try breadboarding both and have a listen.

For the transistor setup, according to the simulation on Falstad, if I change the bottom resistor on the voltage divider to 116.77K I get a centered signal.

Is that valid, or just a simulation glitch?

[wasedadoc]

X3 is the transistor after the the volume control. Its collector is at a suitable dc voltage to attach the base of the emitter follower. There is no need to have any voltage at the exact mid-point.

[cincin]

X3 is the transistor after the the volume control. Its collector is at a suitable dc voltage to attach the base of the emitter follower. There is no need to have any voltage at the exact mid-point.

It all makes a lot more sense now haha, I didn't get that you were referencing the tape deck's circuit.
Good suggestion. Cheers!

[cincin]

X3 is the transistor after the the volume control. Its collector is at a suitable dc voltage to attach the base of the emitter follower. There is no need to have any voltage at the exact mid-point.

Pardon my complete ignorance, but where should the collector of the NPN be connected? I'm assuming the output cap and resistor should be on the emitter side?
Both the opamp and the transistor circuits can operate at various voltages. Is one voltage source preferable to others? Where in the tape deck circuit should I tap for power?
I suppose if I want to use the 4.5V biased signal at the collector of X3, I should find a 9V source? or add a voltage divider for it?
Thanks!

[wasedadoc]

Use the 24v at the right hand end of R27.

[cincin]

Use the 24v at the right hand end of R27.

Base of NPN directly to collector of X3. Use 2k2 emitter resistor. 22uF capacitor, positive end to emitter. 10k resistor from negative end to 0 Volt.

Like this??
I put a 4.5V DC offset on the signal source, like the signal I should find on X3's collector.
The DC supply to the collector of the NPN is 23.5V, as it would be coming from the right hand of R27.
It's not passing signal, so I'm guessing something's wrong.
Thanks again for the help and patience.

[wasedadoc]

Use the 24v at the right hand end of R27.

Base of NPN directly to collector of X3. Use 2k2 emitter resistor. 22uF capacitor, positive end to emitter. 10k resistor from negative end to 0 Volt.

Like this?? It's not passing signal, so I'm guessing something's wrong.
Thanks again for the help and patience.

No. Your simulation does not have the voltage from X3 collector. Simulating an emitter follower at audio frequencies is akin to using a calculator to add 1 and 2. Just switch on the soldering iron and build the real thing.

[ledtester]

...
Like this??
I put a 4.5V DC offset on the signal source, like the signal I should find on X3's collector.
The DC supply to the collector of the NPN is 23.5V, as it would be coming from the right hand of R27.
It's not passing signal, so I'm guessing something's wrong.
Thanks again for the help and patience.

Not sure what's wrong with your Falstad circuit... this one seems to work better:

(Falstad link)

[Xena E]

Simulating an emitter follower at audio frequencies is akin to using a calculator to add 1 and 2. Just switch on the soldering iron and build the real thing.

Refreshing to read. 
+1

E

[Kleinstein]

The signal at the volume pot may not be large enough for line level and one may need a little extra gain, at least have the option for this.
Quality wise the 4558 / TL072 / LF351 should be good enough.

The internal supply is single supply only, but one could add a negative side if needed (similar to the positive side, just the other polarity for the diodes and capacitor).

[wasedadoc]

The signal at the volume pot may not be large enough for line level and one may need a little extra gain, at least have the option for this.
Quality wise the 4558 / TL072 / LF351 should be good enough.

The internal supply is single supply only, but one could add a negative side if needed (similar to the positive side, just the other polarity for the diodes and capacitor).

If higher signal level is needed then an emitter follower will not suffice.  But an opamp does not require a negative supply nor virtual ground.  See schematic.